Streaming of media content over a network may utilize large amounts of resources, both at the content source and in the network. Therefore, point-to-point distribution (i.e., unicasting) tends to not scale well with the size of the target population of consumers. Point-to-multipoint distribution (i.e., multicasting) may be more practical in many cases, as in multicast distribution a single source sends the content to multiple consumers, which allows resource sharing/savings at the source server and all through the network. Multicasting may be used for streaming any suitable type of media content, such as video-on-demand, television, etc.
However, difficulties may be encountered in ensuring a level of communication reliability when distributing media content via multicasting, as a number of end-points involved in coordinating the communication process is larger than for unicast scenarios. One approach for increasing reliability involves utilizing feedback from a consumer to trigger recovery of data. However, such methods may be difficult to implement efficiently due to the amount of computing and communication resources utilized by such feedback mechanisms.